registry / AI-agent control-surface risk / @routerhub/agent-rules
@routerhub/agent-rules
Shared Copilot agent rules and guidelines for RouterHub projects
This report shows the evidence for @routerhub/agent-rules@1.5.247. 63 of 90 known versions are blocked or flagged. See all versions.
AI Security Review
scanned 23d ago · by lpm-firewall-aiThe inspected lifecycle chain performs unconsented install-time mutation of a broad AI-agent control surface, meeting the supplied blocking policy. The personal-file whitelist does not guard the confirmed Copilot write.
Static reason
One or more suspicious static signals were detected.; source matched previously finalized malicious package; routed for review; source fingerprint signature matched known malicious package; routed for review
Trigger
npm installation with lifecycle scripts enabled.
Impact
Package-controlled instructions alter the behavior of agents using the consumer project's Copilot rules.
Mechanism
postinstall.js selects the consumer project and runs merge.js init, which merges package rules and writes .github/copilot-instructions.md without requesting consent.
Policy narrative
The installation hook resolves the consumer project and invokes rule initialization automatically. Initialization writes merged package rules into Copilot instructions regardless of whether personal agent rules are enabled. This crosses into a broad AI-agent control surface without a consent prompt. The hook additionally changes the consumer's package-manager configuration.
AI rationale
The inspected lifecycle chain performs unconsented install-time mutation of a broad AI-agent control surface, meeting the supplied blocking policy. The personal-file whitelist does not guard the confirmed Copilot write.
Evidence
package.jsonpostinstall.jsmerge.js.github/copilot-instructions.md.npmrc
Decision evidence
public snapshotAI identified this as a dangerous AI-agent capability at 99.0% confidence with low false-positive risk.
Evidence for policy risk
- package.json automatically runs postinstall.js after installation.
- postinstall.js locates the consumer project and launches merge.js init there without a consent prompt.
- merge.js initialization calls mergeAgents, which writes consumer Copilot instructions outside the personal-file whitelist gate.
- postinstall.js also appends package-manager settings to the consumer project's .npmrc.
Evidence against
- Personal AGENTS.md and CLAUDE.md generation is gated by a local identity whitelist or environment override; this does not gate Copilot instructions.
Behavioral surface
ChildProcessDynamicRequireEnvironmentVarsFilesystemShell
HighEntropyStringsUrlStrings
Source & flagged code
9 flagged · loading sourcepackage.jsonView file
•scripts.postinstall = node postinstall.js
High
Install Time Lifecycle Scripts
Package defines install-time lifecycle scripts.
package.jsonView on unpkg•scripts.postinstall = node postinstall.js
Medium
Ambiguous Install Lifecycle Script
Install-time lifecycle script is not statically allowlisted and needs review.
package.jsonView on unpkgpostinstall.jsView file
•matchType = normalized_sha256
matchedPackage = @routerhub/agent-rules@1.5.245
matchedPath = postinstall.js
matchedIdentity = npm:QHJvdXRlcmh1Yi9hZ2VudC1ydWxlcw:1.5.245
similarity = 1.000
summary = normalized source hash matched finalized malicious source
High
Known Malware Source Similarity
Source file is highly similar to a previously finalized malicious package; route for source-aware review.
postinstall.jsFindings
6 High5 Medium4 Low
HighInstall Time Lifecycle Scriptspackage.json
HighSemantic Analysis Limitedscripts/agent-browser-cleanup.sh
HighKnown Malware Source Similaritymerge.js
HighKnown Malware Source Similaritypostinstall.js
HighKnown Malware Source Similarityskills/screenshot-annotate/annotate.cjs
HighKnown Malware Source Fingerprint Signaturemerge.js
MediumAmbiguous Install Lifecycle Scriptpackage.json
MediumDynamic Requirepostinstall.js
MediumEnvironment Vars
MediumShips Build Helperscripts/agent-browser-cleanup.sh
MediumStructural Risk Force Deep Review
LowScripts Present
Low
Affected versions and remediation
This report applies to @routerhub/agent-rules@1.5.247.
The versions list below shows the verdict for each recorded version.
- Avoid installing @routerhub/agent-rules@1.5.247. Remove it from direct dependencies and check your lockfile for transitive copies.
- Choose an independently verified alternative or release. This report does not establish that other versions are safe.
- If this version ran, investigate the affected machine and build environment. Rotate credentials it could access and rebuild from a trusted environment.
Evidence last updated: .
Think this verdict is wrong? Dispute it on GitHub or read how reports are produced.
Versions
90 known versions1.5.247latestshownMaliciousSoft block: AI-agent control surfaceBlockPublishedSep 30, 2026, 9:00 AM1.5.245MaliciousSoft block: AI-agent control surfaceBlockPublishedSep 29, 2026, 2:10 AM1.5.244MaliciousSoft block: AI-agent control surfaceBlockPublishedSep 28, 2026, 10:03 AM1.5.243MaliciousSoft block: AI-agent control surfaceBlock